March 5th, 2008 by David Poteet

Lessons learned from Imperial College London

Imperial College London web team: From left to right, Sian Berry, Pamela Michael, Caroline Davis, Becky Manning, Peter Gillings, Tom Johnson, Maurice Farmer, Tom Miller, Saul Batzofin and Stuart Croghan

Jeannine Everhart and I traveled to London in early February to meet with Imperial College London for an end-of-project review. We met with the entire web team and went through all aspects of the project including feedback from students and staff, analytics, review of project objectives and a discussion of lessons learned.

Overall, what worked particularly well on this project?

Working across the pond – Imperial commented that the working relationship between themselves and NCM was really successful. Initial concerns with working with an agency across the pond and in a completely different time zone were quickly forgotten as easy communications were established – regular phone calls, instant messenger and basecamp project software. In fact, the distance became a real benefit to the project as it helped focus the Imperial web team to ensure that during NCM visits to London, everyone committed the time to really concentrate on developing the project.
Project oversight and approvals – Imperial has a web governance board that oversees web policy. This group gave authority to Tom Miller and his team to run the project and make most of the decisions. The fact that the board trusted Tom with these decisions made the process much smoother.
Audience research – the research and mental models that we did prior to starting information architecture helped not only guide our architectural decisions, but also helped answer questions that various stakeholders would ask. According to Tom Miller, "that [the audience research] was the ticket that bought us the credibility and led to the smoothness later on."
Communication among different groups on campus – a huge win for the project was more intangible. The web team was comprised of people from different groups on campus that don't usually get to work together. During the project they were able to communicate and understand each other's roles much better, and it has fostered connections that will bring benefits for a long time.
Workstream teams – Imperial set up sub-teams to handle various aspects or "workstreams" of the project, like design, content, implementation etc. Each team had a designated leader. This helped Pamela Michael, the project manager, cover a lot more ground.

What things could we have done better?

Critique designs for functionality, not just aesthetics – there were a few places where design concepts had what were essentially product ideas embedded within them. The project manager at Imperial pointed out that she should have asked people to comment not only on the visual design, but also the feasibility of implementing some of these ideas in time for the initial launch.
A way to see all versions of a particular design – because we were designing so many different landing pages and templates, it got a bit confusing to see all the evolutionary steps a page went through to get to the final version. We had 2 different web sites where people could view mockups and various documents, and this contributed to the confusion.
Design spilled over too much into implementation – as often happens, we were still designing some parts of the site as others were being built. We allowed this to go on too long, and it put undue pressure on the implementation team.
Implementation should have been planned out more thoroughly – this phase of the project was handled by Imperial staff, so we (at NCM) didn't put as much thought into breaking out all the steps of the process. It would have helped if we both had spent more time up front thinking through this phase. Still, Imperial did an excellent job of putting it all together.
Workstreams could have started earlier – Imperial indicated it would have helped to get them organized earlier in the process. Still, this sub-team concept was very successful.
Initial goals could have been more measurable – In hindsight Imperial realized that some of the goals they set for the project could have been defined in a way that would make it easier to measure whether they had been achieved or not.
Should have set up some benchmark analytics before launching the new site – The analytics in use for the existing site were not very consistent, and in some cases did not capture all of the areas of the site. We wanted a way to make a fair comparison between the old and new sites, and this was not easy.

Feedback from a visitor survey was largely positive. Over 80% of respondents said they were able to complete all or some of their most recent task on the web site. Written comments covered the entire spectrum. Here are some examples (including both good and bad):

"This is a superb website. It puts Imperial ahead of its peers (i.e. Oxford, Cambridge, MIT, Yale, Harvard). I love the look and feel. It is fresh, modern and information-rich without being fussy. The compelling images and imaginative use of colour instantly "grabbed" me.
"The site fails miserably to project Imperial's image as a world class institute. It is tacky and poorly designed and reminds me of late 20th century, run-of-the-mill websites. It would be more suitable for a cinema/theatre than an institute of learning, science and research." [This one's my favorite - DP]
"I think it's a much easier website to navigate than the old one and much more colourful and inviting."
"Crap"
"Looks fantastic, love the new Staff page! Well done to all of you."
"Difficult to navigate. Very small text. The main staff page is simply too busy."
"Which American college's website did you copy? It is indeed way better, but shockingly similar to US colleges."
"It is now more difficult to get to research and IT pages."
"Nice layout but a lot of the information needs to be updated for it to actually be of any use."
"Content: Neither better or worse than the previous site. Ultimately, quality relies on the individual faculty members', programmes, departments' and divisions' pages. VERY little deep content has been revised."
"I think the new site is a massive improvement. It's about as good as any university website, certainly in the UK."
"Absolutely fantastic. Haven't really poked around too much yet, but looks brilliant. 5/5 from me! Well done guys! Long overdue, but well worth the wait."

Overall feedback about the design and information architecture was very positive. There were several comments about site content, and the fact that the redesign had not really addressed problems with deeper content. Those comments were entirely valid.

Why is content still an issue?

The redesign project did not address deeper content at all, so in most cases existing materials just moved to a new template without edits or updates. This is really part of a broader issue, and something that has implications for the way the Imperial community handles web communication in general.

Here are some factors that the Imperial team pointed out:

There are currently over 300 web site owners at Imperial, who have responsibility for a page or a collection of pages. Many of these owners are personal assistants or people for whom the web site is only a small part of their regular job responsibilities (sound familiar?). Their supervisors may not place much importance on the web site.
No one is looking at the site, or even sections of the site, as a whole user experience. People are focused more on individual pages rather than the relationships and FLOW between them.
College policy states that every page should be reviewed at least once a year to determine if it is accurate and current, but this is not enforced. As a result many pages have gone years without review.

Possible Solutions:

The Imperial web team had several ideas for improving site content.

Recognize that it's a cultural issue. It won't change until top managers place greater importance on this communication role, and people in the role perceive that importance.
Address importance of this role with directors, heads of various groups on campus.
Create a role of "information manager," someone responsible for guiding the management of content across the campus.
Bring site owners together for learning, sharing best practices.

An end-of-project review is one of those things that everyone thinks is a good idea but few take the time to do. It was a great experience for us to meet with the entire team from Imperial and take such a thorough look at the process and results. And dinner afterwards at The Oratory was fantastic.

But then there was that strange journey to the loo…

Read a case study on the Imperial College London website redesign project.